Two of the most closely watched pipeline readouts in obesity pharmacology are now in. In February 2026, Novo Nordisk reported REDEFINE 4, the first head-to-head trial between CagriSema and tirzepatide. In May, Eli Lilly reported TRIUMPH-1, the pivotal phase 3 for retatrutide. For the first time there is enough data to draw a real competitive map - not built on cross-trial guesses, but on actual results.

This post scores each drug on where it stands as of May 2026. For context on why the absolute percentages require adjustment before comparing across trials, see the first post in this series.

The Approved Baseline #

Two drugs are currently approved for chronic weight management in non-diabetic adults with obesity.

Semaglutide 2.4 mg (Wegovy, Novo Nordisk) - approved by the FDA in June 2021. STEP 1 (NCT03548935, 68 weeks) showed a treatment-regimen weight loss of 14.9%. In SURMOUNT-5, the only published head-to-head against tirzepatide, semaglutide achieved -13.7% against tirzepatide's -20.2%, a 6.5 percentage point gap under identical trial conditions.

Tirzepatide 15 mg (Zepbound, Eli Lilly) - approved by the FDA in November 2023. SURMOUNT-1 (NCT04184622, 72 weeks) showed -20.9%. In SURMOUNT-5, -20.2%. In REDEFINE 4, a head-to-head against CagriSema over 84 weeks, tirzepatide achieved -25.5%.

Oral semaglutide 25 mg (Wegovy pill, Novo Nordisk) - approved by the FDA in January 2026. OASIS 4 (NCT05564117, 64-week primary endpoint, n=205) showed -13.6% weight loss on the treatment-policy estimand and -16.6% on the trial-product (on-treatment) estimand versus -2.2% on placebo. Total gastrointestinal adverse events were reported in 74.0% of participants versus 42.2% on placebo. The trial enrolled a small sample across 22 sites in four countries, which limits the precision of the estimates. In a manufacturer-sponsored indirect treatment comparison (ORION study, Obesity Medicine Association 2026), oral semaglutide 25 mg was associated with significantly greater mean weight loss than orforglipron 17.2 mg using data from OASIS 4 and ATTAIN-1; this is an adjusted indirect comparison, not a head-to-head RCT.

These are the benchmarks everything else gets measured against.

The Challengers #

CagriSema (Novo Nordisk) #

CagriSema is a fixed-dose subcutaneous co-formulation of cagrilintide (an amylin analogue) and semaglutide 2.4 mg. The pivotal REDEFINE 1 trial showed -20.4% weight loss at 68 weeks on the treatment-policy (intent-to-treat) primary endpoint (NEJM, June 2025); the on-treatment efficacy estimand was -22.7%. On the efficacy estimand, CagriSema looked modestly ahead of tirzepatide's SURMOUNT-1 result (-20.9%, treatment-regimen) on a cross-trial read, which led to a widespread expectation that CagriSema would prove superior in a head-to-head.

That expectation was tested in REDEFINE 4 (NCT06131437, 84 weeks, open-label, n=809). Tirzepatide achieved -25.5%; CagriSema -23.0%. CagriSema failed to meet the pre-specified non-inferiority threshold. The drug that looked ahead on cross-trial data lost the direct comparison by 2.5 percentage points. This is the cleanest illustration yet of the problem described in the first post: cross-trial rankings are directional at best.

Novo Nordisk has submitted CagriSema for FDA approval. The regulatory submission is intact - REDEFINE 4 was a non-inferiority failure, not a safety finding - but the commercial narrative around CagriSema changed substantially on that result.

Retatrutide 12 mg (Eli Lilly) #

Retatrutide is a triple agonist targeting the GLP-1, GIP, and glucagon receptors. TRIUMPH-1 (NCT05929066, 80 weeks) reported on 21 May 2026: -25.0% on the treatment-regimen estimand and -28.3% on the efficacy estimand. Both are the highest figures reported from a pivotal obesity trial.

One design note matters. TRIUMPH-1 excluded type 2 diabetes by diagnosis, without requiring an HbA1c measurement at screening. This is the same approach used in SCALE Obesity (liraglutide, 2015), and different from STEP and SURMOUNT, which measured HbA1c at screening and excluded anyone at or above 6.5%. The practical consequence is that TRIUMPH-1 may have enrolled participants with undiagnosed prediabetes, which could push the headline number above what a STEP-style protocol would produce. There is no head-to-head trial between retatrutide and tirzepatide yet, so the true pharmacological gap is unknown.

Lilly has indicated it expects to file for FDA approval in 2026. An extended follow-up of participants with a BMI at or above 35 showed -30.3% weight loss at 104 weeks. At 12 mg, 62.5% of patients lost at least 25% of their body weight, 45.3% at least 30%, and 27.2% at least 35%. On safety, 11.3% of patients on 12 mg discontinued due to adverse events compared with 4.9% on placebo; dysesthesia was reported in 12.5% of 12 mg patients versus 0.9% on placebo. All figures from the Lilly press release of 21 May 2026.

Horizontal bar chart showing pivotal obesity trial results for liraglutide (-8.4%), semaglutide (-14.9%), tirzepatide (-20.9%), and retatrutide (-25.0% treatment-regimen, -28.3% efficacy estimand, -30.3% at 104 weeks in BMI ≥35 subgroup). Retatrutide is the top bar in cyan with two faded extensions.
Pivotal obesity trial results by drug. Solid bars show the treatment-regimen estimand. The first faded extension on retatrutide shows the efficacy estimand (-28.3% at 80 weeks); the lighter extension shows the 104-week result in the BMI ≥35 subgroup (-30.3%). Trial durations differ: 56 weeks (SCALE) to 80 weeks (TRIUMPH-1). Source: Lilly press release, 21 May 2026.
Grouped bar chart showing the percentage of patients achieving at least 25%, 30%, and 35% body weight loss with retatrutide 4 mg, 9 mg, and 12 mg versus placebo in TRIUMPH-1. At 12 mg: 62% at ≥25%, 45% at ≥30%, 27% at ≥35%.
Responder analysis from TRIUMPH-1 (efficacy estimand, 80-week primary endpoint). At the 12 mg filing dose, 62.5% of patients lost at least 25% of body weight, 45.3% at least 30%, and 27.2% at least 35%. The dose-response gradient is consistent across all three thresholds. Source: Lilly press release, 21 May 2026.

Amycretin (Novo Nordisk) #

Amycretin is a GLP-1 and amylin co-agonist in a single molecule. A Phase 1 first-in-human study on the oral formulation showed 13.1% weight loss at 12 weeks at the highest dose tested, versus 1.2% on placebo - an early, short-duration signal that cannot be compared directly to 68-80 week pivotal results. Phase 3 NEXUS trials are ongoing. No pivotal data is available and Amycretin is the most uncertain near-term asset in the field.

Orforglipron (Eli Lilly) #

Orforglipron (Foundayo) is a once-daily oral non-peptide GLP-1 agonist. ATTAIN-1 (NEJM, November 2025, n=3,127, 72 weeks) showed -7.5%, -8.4%, and -11.2% at 5.5 mg, 9 mg, and 17.2 mg respectively, versus -2.1% on placebo. At the highest dose, 36% of patients lost at least 15% of body weight. The drug was approved by the FDA in April 2026 for chronic weight management - the first oral GLP-1 agonist to receive that indication. At -11.2%, it falls below the injectable benchmarks but proves the access argument: a pill that delivers that level of weight loss is sufficient for the large population of patients who are injection-averse or in markets where injectables are unaffordable or poorly reimbursed. Separately, for type 2 diabetes, ACHIEVE-3 (The Lancet, February 2026) showed superior HbA1c reduction versus oral semaglutide; a T2D regulatory filing is planned for 2026.

Grouped horizontal bar chart showing adverse event rates in TRIUMPH-1 by dose. Nausea 42.4% at 12 mg vs 14.8% placebo. Diarrhea 32.0% at 12 mg vs 13.5%. Constipation 26.1% vs 10.9%. Vomiting 25.3% vs 4.8%. Dysesthesia 12.5% vs 0.9%. Discontinuation due to adverse events 11.3% vs 4.9%.
Adverse events in TRIUMPH-1 by dose. Gastrointestinal events follow the pattern seen with other GLP-1 and GIP/GLP-1 agonists. Dysesthesia (abnormal skin sensation) is a distinguishing feature of retatrutide not commonly reported in other incretin trials; it occurred in 12.5% of 12 mg patients vs 0.9% on placebo and was generally mild to moderate. Discontinuation due to adverse events was 11.3% at 12 mg vs 4.9% on placebo. Source: Lilly press release, 21 May 2026.

Adverse events across trials #

Comparing tolerability across trials is harder than comparing efficacy - not every paper breaks out individual gastrointestinal events in the main table. The chart below draws on main-table AE data for semaglutide (STEP 1), tirzepatide (SURMOUNT-1), and oral semaglutide 25 mg (OASIS 4), and on the published supplementary appendices (Table S7) for CagriSema (REDEFINE 1, NEJM 2025) and orforglipron (ATTAIN-1, NEJM 2025). TRIUMPH-1 data is from the Lilly press release; the full journal paper is expected at ADA June 2026.

Grouped horizontal bar chart comparing adverse event rates across six pivotal obesity trials. CagriSema has the highest nausea rate at 55.0%. Tirzepatide has the lowest GI event rates across all categories. Retatrutide is the only drug showing dysesthesia at 12.5% and has the highest discontinuation due to adverse events at 11.3%. Oral semaglutide 25mg and injectable semaglutide 2.4mg have similar nausea rates but oral sema has notably lower diarrhea.
Adverse events at the top approved or filing dose in each pivotal trial. Tirzepatide has the most favourable GI profile. CagriSema's high nausea rate (55.0%, REDEFINE 1 supplement Table S7) reflects additive effects of semaglutide and cagrilintide. Oral semaglutide 25 mg and injectable semaglutide 2.4 mg have near-identical nausea rates (46.6% vs 44.2%), but oral sema has markedly lower diarrhea (17.6% vs 31.5%), possibly reflecting different routes of absorption. Dysesthesia appears unique to retatrutide and was not reported in the other trials. Despite the highest nausea rate, CagriSema had the lowest discontinuation due to AE (5.9%), suggesting most events were transient. Sources: STEP 1 · SURMOUNT-1 · OASIS 4 · REDEFINE 1 · ATTAIN-1 · Lilly press release 2026-05-21.

The Scorecard #

Horizontal bar chart showing pivotal trial results (ghost bars) versus head-to-head results (solid bars) for semaglutide, CagriSema, tirzepatide, and retatrutide. Tirzepatide's head-to-head result of -25.5% in REDEFINE 4 exceeds its own pivotal. Retatrutide shows -25.0% in TRIUMPH-1 with no head-to-head yet.
Ghost bars: each drug's own pivotal trial result. Solid bars: head-to-head result where available. CagriSema and tirzepatide faced each other in REDEFINE 4; semaglutide and tirzepatide in SURMOUNT-5. Retatrutide has no head-to-head trial yet. REDEFINE 1 ghost bar shows the treatment-policy (ITT) primary endpoint from NEJM June 2025 (-20.4%); on-treatment efficacy estimand was -22.7%. REDEFINE 4 result from Novo Nordisk press release, 23 Feb 2026 (efficacy estimand).

The chart above shows a consistent pattern across both completed head-to-head trials: the drug that looked ahead on cross-trial data did not win the direct comparison. CagriSema looked ahead of tirzepatide on an efficacy-estimand cross-trial read (REDEFINE 1 -22.7% vs SURMOUNT-1 -20.9%); tirzepatide won the direct comparison by 2.5 points. Semaglutide in SURMOUNT-5 came in 1.2 points below its own STEP 1 result - same drug, same dose, different trial, different number. That gap is not noise. It is the population variability that cross-trial comparisons paper over.

Drug Company Route Status Pivotal result H2H result Notes
Semaglutide 2.4 mg (Wegovy) Novo Nordisk Weekly SC Approved (2021) -14.9% (STEP 1, 68 wks)
PMID 33567185		 -13.7% vs tirze (SURMOUNT-5)
PMID 40353578		 Lost SURMOUNT-5 by 6.5pp
Tirzepatide 15 mg (Zepbound) Eli Lilly Weekly SC Approved (2023) -20.9% (SURMOUNT-1, 72 wks)
PMID 35658024		 -20.2% vs sema (SURMOUNT-5)
PMID 40353578
-25.5% vs CagriSema (REDEFINE 4)
Novo Nordisk press release		 Won both h2h trials to date
CagriSema 2.4/2.4 mg Novo Nordisk Weekly SC Filed (2026) -20.4% TR / -22.7% eff (REDEFINE 1, 68 wks)
NEJM, June 2025		 -23.0% vs tirze (REDEFINE 4)
Novo Nordisk press release		 Failed non-inferiority vs tirze; open-label
Retatrutide 12 mg Eli Lilly Weekly SC Filing (2026) -25.0% TR / -28.3% eff (TRIUMPH-1, 80 wks)
Lilly press release		 No h2h trial yet No HbA1c screening cap (SCALE-era design); triple agonist (GLP-1/GIP/GCG)
Amycretin Novo Nordisk Oral / SC Phase 3 No pivotal data No h2h data ~13% at 12 wks (Ph1, oral); NEXUS Ph3 trials ongoing
Orforglipron (Foundayo) Eli Lilly Oral daily Approved (obesity, Apr 2026) -11.2% at 17.2 mg (ATTAIN-1, 72 wks)
NEJM, Nov 2025		 No h2h obesity data First oral GLP-1 approved for obesity; ACHIEVE-3 T2D trial (Lancet, Feb 2026); T2D filing planned 2026
Oral semaglutide 25 mg (Wegovy pill) Novo Nordisk Oral daily Approved (obesity, Jan 2026) -13.6% TP / -16.6% eff (OASIS 4, 64 wks, n=205)
NEJM, Sep 2025		 Indirect comparison vs orforglipron (ORION ITC, not RCT) Same molecule as Wegovy sc; small trial (n=205+102 at 22 sites); GI AEs 74.0% vs 42.2% placebo
TR = treatment-regimen estimand. Eff = efficacy estimand. SC = subcutaneous injection. H2H results are the drug's own arm in the listed trial. Bold winning arms shown in text above. Sources: ClinicalTrials.gov, published papers, and press releases as noted.

What the Numbers Still Don't Tell You #

Even with two head-to-head trials now complete, cross-trial comparisons between the remaining drugs remain unreliable. Retatrutide's treatment-regimen result of -25.0% vs tirzepatide's -20.9% looks like a 4-point gap. The actual gap - if the same patients went through identical protocols - is unknown. The only numbers in this space that need no asterisk are from head-to-head trials under matched conditions: SURMOUNT-5 (tirzepatide -20.2% vs semaglutide -13.7%) and REDEFINE 4 (tirzepatide -25.5% vs CagriSema -23.0%). Everything else is a starting point for analysis, not a conclusion.

The Investment Angle #

Eli Lilly enters the second half of 2026 with two approved obesity drugs and a third (retatrutide) in late-stage regulatory review. The clinical data is running in one direction: tirzepatide won SURMOUNT-5 against semaglutide and REDEFINE 4 against CagriSema. Retatrutide reported the largest pivotal obesity trial number ever recorded. Whether that translates into market share is a separate question - distribution, pricing, reimbursement, and supply all matter more in the short run than a percentage difference in a 500-person trial - but the clinical narrative is clearly Lilly's.

Novo Nordisk retains dominant commercial scale, an established prescriber base, and a manufacturing infrastructure that competitors cannot replicate quickly. CagriSema's regulatory submission is intact; approval is likely even after REDEFINE 4. Amycretin is still in play, and if it proves competitive with injectables in an oral formulation, the market expands rather than contracts. But the picture as of mid-2026 is that every clinical comparison between a Lilly drug and a Novo Nordisk drug - SURMOUNT-5 and REDEFINE 4 - has gone to Lilly.

The oral efficacy question is already partially answered. Orforglipron at -11.2% (highest dose, ATTAIN-1) is below semaglutide's -14.9% mean, but the FDA approved it anyway and the drug is on the market since April 2026. That result defined one segment: injection-averse patients and price-sensitive markets where a pill at 11% is enough. What it did not do is close the gap to injectables. The remaining open question is Amycretin: if Phase 3 NEXUS produces a number above 15%, the oral category shifts from an access story to an efficacy story - and that is where the addressable population roughly doubles.